RE: Modelling VRA core

Mark,

I prefer the "hedgehog" style.

Qualifiers in DC are expressed as subproperties.  For example from the DC
terms schema:

	dcterms:alternative rdfs:subPropertyOf dc:title

Might that be a way to capture:

vracore extract:
TITLE
Qualifiers:
    Title.Variant
    Title.Translation
    Title.Series
    Title.Larger Entity
Definition: ...
Dublin Core:  TITLE

is

vra:titleVariant  rdfs:subPropertyOf  vra:title .
vra:title         rdfs:subPropertyOf  dc:title .

This would lead to flat data struture per record and by quite like the
exmple they give of key/value lines.

There might also be a need to process the data to add more structrure.
However, this will depend on the quality of the input, any controlled
vocabulary used and not just the schema.  Your example has a good case:

    Measurements.Dimensions = 24.5 x 35 cm    // cm only once

yet the next example they give is:

   Measurements.Dimensions = 72 dpi           // dpi? Its not a "dimension"
really

so the approach I prefer is to capture the non-RDF metadata as completely as
possible in RDF.  Here, that means each value is an unprocessed string.
Then, later processing can be applied to put in the structure but that is on
a per-dataset basis, not for all data using the VRAcore schema.


Technique = etching
Technique = drypoint

Could use a repeated property here, not a bag.  Number of reasons: quite a
few properties might be repeated, but often will not (e.g. title).  Using a
bag makes the case of one instance different from two.  Second, of the same
property/value pair appears, it is not adding more information so the fact
that only one statement gets added seems natural.

I also like camel naming - properties are lowercase, with subsequent words
capitalized.  Actaully, I'm not that keen on it but that is the more common
way of doing it.  I grew up using a_b_c style but have found that rather
rare these days.

	Andy

-----Original Message-----
From: Butler, Mark [mailto:Mark_Butler@hplb.hpl.hp.com] 
Sent: 15 August 2003 17:12
To: 'SIMILE public list'
Subject: Modelling VRA core



Hi team

I have been having a go at writing an RDFS Schema for VRA Core because one
does not exist at the moment. To test this out, I've been creating some
example instance data. However when we model things we often have to decide
whether to structure the data e.g. group some properties together or whether
to leave the data unstructured (referred to as a "hedgehog model"). I'm
trying to work through the pros and cons of each approach, but I'd be
interested to hear other people's opinions on this. I wonder if qualifiers,
as used in DC or VRA, imply that the vocabulary authors would prefer
structure in the data - what do others think?

Here are two example RDF instance data files to illustrate the difference:

First the structured version:

<?xml version="1.0"?>
<!DOCTYPE rdf:RDF [
        <!ENTITY rdfns 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'>
        <!ENTITY vracorens 'http://www.vraweb.org/vracore3/20030915#'>
   ]>
<rdf:RDF xmlns:rdf="&rdfns;"
         xmlns:vra="&vracorens;"
         xmlns="&vracorens;">
<rdf:Description>
<recordType rdf:resource="&vracorens;Work"/>
<type>print</type>
<title rdf:parseType="Resource">
	<theTitle>This is how it happened</theTitle>
	<variant>As Sucedi</variant>
</title>
<measurements rdf:parseType="Resource">
	<dimensions>24.5 x 35 cm</dimensions>
</measurements>
<material rdf:parseType="Resource">
	<medium>ink</medium>
	<support>paper</support>
</material>
<technique>
	<rdf:Bag>
		<rdf:li>etching</rdf:li>
		<rdf:li>drypoint</rdf:li>
	</rdf:Bag>
</technique>
<creator rdf:parseType="Resource">
	<personal_name>Francisco Jose de Goya y Lucientes</personal_name>
	<role_name>printmaker</role_name>
</creator>
<date_creation>ca. 1810-1814</date_creation>
<location rdf:parseType="Resource">
	<current_repository>Ann Arbor (MI,USA), University of Michigan
Museum of Art</current_repository>
	<creation_site>Madrid (ESP)</creation_site>
</location>
<idNumber rdf:parseType="Resource">
	<current_accession>1977/2.15</current_accession>
</idNumber>
<stylePeriod rdf:parseType="Resource">
	<theStylePeriod>Romanticism</theStylePeriod>
</stylePeriod>
<culture>Spanish</culture>
<subject>war</subject>
<relation rdf:parseType="Resource">
	<partOf>Part of Disasters of war</partOf>
</relation>
<description>
This is how it happened is No. 47 (33) from the series "The 
Disasters of War", 4th edition, plates for the series ca. 1810-14, 1820, 4th
edition was published 1906.
</description>
<rights>Weber family trust</rights>
</rdf:Description>
</rdf:RDF>


then the "hedgehog" version

<?xml version="1.0"?>
<!DOCTYPE rdf:RDF [
        <!ENTITY rdfns 'http://www.w3.org/1999/02/22-rdf-syntax-ns#'>
        <!ENTITY vracorens 'http://www.vraweb.org/vracore3/20030915#'>
   ]>
<rdf:RDF xmlns:rdf="&rdfns;"
         xmlns:vra="&vracorens;"
         xmlns="&vracorens;">
<rdf:Description>
<recordType rdf:resource="&vracorens;Work"/>
<type>print</type>
<title>This is how it happened</title>
<title_variant>As Sucedi</title_variant>
<measurements_dimensions>24.5 x 35 cm</measurements_dimensions>
<material_medium>ink</material_medium>
<material_support>paper</material_support>
<technique>
	<rdf:Bag>
		<rdf:li>etching</rdf:li>
		<rdf:li>drypoint</rdf:li>
	</rdf:Bag>
</technique>
<creator_personal_name>Francisco Jose de Goya y
Lucientes</creator_personal_name>
<creator_role_name>printmaker</creator_role_name>
<date_creation>ca. 1810-1814</date_creation>
<location_current_repository>Ann Arbor (MI,USA), University of Michigan
Museum of Art</location_current_repository>
<location_creation_site>Madrid (ESP)</location_creation_site>
<idNumber_current_accession>1977/2.15</idNumber_current_accession>
<stylePeriod>Romanticism</stylePeriod>
<culture>Spanish</culture>
<subject>war</subject>
<relation_partOf>Part of Disasters of war</relation_partOf>
<description>
This is how it happened is No. 47 (33) from the series "The 
Disasters of War", 4th edition, plates for the series ca. 1810-14, 1820, 4th
edition was published 1906.
</description>
<rights>Weber family trust</rights>
</rdf:Description>
</rdf:RDF>

any comments?

regards

Dr Mark H. Butler
Research Scientist                HP Labs Bristol
mark-h_butler@hp.com
Internet: http://www-uk.hpl.hp.com/people/marbut/

Received on Monday, 18 August 2003 05:00:10 UTC